[pnrailway]

I think we are getting away from what this thread started out about but my "cheep" computer has a 2.33 quad core
6 gb of ram
came with onboard video but I added a
nVidia 9800 GT w/ 500mb onboard ram
(2) internal 1 tb hard drives
plus (3) external drives
(1) 1 tb
(2) 500 gb

Now back to your regular sponsor

[thegrindre]

:oldstry: We have gone astray, haven't we?

:pardon:

[Genma Saotome]

thegrindre, on 25 September 2012 - 09:19 AM, said:

:oldstry: We have gone astray, haven't we?

:pardon:

Not to worry, I've split this off into its own thread.

IMO the big difference between any PC good enough for MSTS and one that is good enough for Open Rails is MSTS doesn't need multi-core CPU's and while neither does OR it will take advantage of them if they're present. IOW, most any multi-core PC running Open Rails should outperform MSTS when using the same route & equipment. A quad core CPU will be somewhat better than any dual core CPU, all other things equal. Toss in Win7 and the differences should increase to OR's advantage (Win7 just does multi-core/multi-thread processing better than any older OS from Microsoft). That does not mean you need a quad core PC running WIN7 to run OR, just that upgrading to newer hardware will improve OR performance much more than it will improve MSTS performance.

Also, different routes will place different demands on the machine... there are some that don't need anywhere near as much CPU horsepower as others. One such example would be the original Tehachippi route vs. the later version that uses Scalerail; SR simply needs more CPU cycles than does Xtraks.

Bottom line is it is hard for us to compare machine performance because too few of us have the same favorite routes and equipment to do head to head comparisons of PC's. About all that can be certain is OR should outperform MSTS on newer hardware because MSTS doesn't know how to take advantage of newer hardware and OR does.

[Genma Saotome]

Fred, on 25 September 2012 - 04:45 AM, said:

I am assuming that V1.0 will be the final release and V1.1 will have the bugs worked out. :oldstry:

That is not a good assumption Fred. AFAIK V1.0 will have all of the features that are in MSTS for which the team feels motivated to write the code. It could be 100% equal to MSTS... or something less (e.g., consider the example of moving water in MSTS... if nobody wants to code that and instead turns their attention to adding new features to signalling, is that is it a terrible loss?).

IOW, V1.0 has nothing whatsoever to do with ending the project and everything to do with finishing the compatibility phase of creating a new train simulator.

As far as waiting goes, sure, for everything working perfectly you probably should wait. But not taking a "looksee" right now, you're missing out on some pretty impressive work.

[Lindsayts]

Fred, on 25 September 2012 - 04:45 AM, said:

As I said in another thread several weeks ago, OR is THE train simulator of the future, it's just that its day hasn't come yet. Working in the computer industry for many years I have developed an aversion to unfinished software that is kicked out the front door simply because a deadline has been reached. I am assuming that V1.0 will be the final release and V1.1 will have the bugs worked out. :oldstry:

You have not understood the development process used in open source programs

Openrails is being developed open source. The developers are all volunteers who all have jobs and families and people who depend on them. The reason they are working on the trainsim is they wish for a GOOD trainsim and willing to put in the hard yards to make it happen. Such development almost NEVER works to deadlines but releases occur when the developers think they have something worth releasing. As they themselves are looking for a decent train sim the standards are usually high. Usually in open source development when a release occurs the developers are looking to the community to test the release throughly to see how it works "in the wild". This is done as it is very difficult for a small group to test everything as well as they would like it to be tested.

Users testing a release is one of the major advantages of opensource development and if one can one should test releases as they come up and if one can submit a well thought out performance/bug report. This way bugs are found sooner and the developers unlike in most commercial programs WILL BE LISTENING. There has been plenty of evidence of this in OR over the time span of its existance.

I am an open source developer having worked in the past on the parrallel port and the framebuffer graphics drivers in the LInux kernel. Currently working on a terrain editing program, this being in fact my first "user" program all previous work being on the kernels and libraries. Nice to get on to something that does not usually crash when one has done something strange.

Lindsay

[Lindsayts]

I have run OR on a variety of systems and in fact still do, the lowest spec system being around 5 years old. This being a Gigabyte P35 SO3 motherboard originally with 2 gig of ram a dual core 2.4 ghz processor and a Radeon 3870 graphics card. This system currently has a 4 core 3 ghz proc, a radeon 4870 GPU and 4 gig of ram.

The highest spec system is an ASUS P6X58D Premium MB with an i7 950 and a NVidia GTX 570 (I think) and 12 gig or ram. Both systems are running Windows XP pro SP3.

All systems in all configurations have run OR well although naturally the performance is better on the latter machines. Particularly on the 4 core processors.

Lindsay

[Matej Pacha]

Just a note:
If anybody has problems with low FPS (under 15), please don't use the advanced adhesion model - use in-game key combination "Ctrl+Alt+X" or the old menu Options/Simulation->Advanced adhesion model. This model requires smaller time steps = higher FPS. We are using time step dividing but it cannot be used on slow computers due to high CPU load.

Matej

[Lindsayts]

Eldorado.Railroad, on 03 October 2012 - 08:37 AM, said:

Please consult this before going to the two links below:

If you have the money...

why not spend it?

The specs are:
Intel Core i7-980X Extreme Edition w/Corsair H100, ASUS Rampage III Extreme, Corsair DOMINATOR-GT 6GB DDR3 1600, GeForce GTX 680, X-Fi Titanium Fatal1ty, Corsair Force Series GT 180GB SSD, WD VelociRaptor 600GB, Corsair AX1200, Corsair 800D, Windows 7 Ultimate 64/SP1

I am sure I could get those FPS down even on that machine, with some killer routes.

Maybe Santa will come by early this year?

Eldorado

Hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm.............................

Actually one can do better than that these days for far less money
My latest system has a Gigabyte X79-UD5 MB with and i7 3920, Graphics is a NVidia GTX 570 Super Overclocked.
The MB and CPU both were less than the cost of the i7 980 by itself, the main differnce with the CPU's being 4 core V 6 core.

At this stage OR cannot effectively use a 6 core system so the great extra cost of a 6 core system is worthless, also a point worth considering is a train sim is VERY object intensive. This uses a large amount of CPU time and memory bandwidth to sort all these objects out but essentailly only 2 cpus can be used. Another point to consider is it appears the current crop of GPU's are real good at texturing objects as opposed to rendering a large number of objects. These points mean there is a good chance that a real high specs system may not be the most cost effective solution for OR.

The Gigabyte X79,the i7 3920 and the GTX 570 were purchased specficly for OR with the above points in mind as a cost effective solution for a well performing system, the GTX 570 SOC particularly being good value for money at the time of purchase.

Lindsay